A Los Angeles Times report published Tuesday says the U.S. semiconductor buildout could face a skilled-worker deficit of as much as 157,000 full-time workers by 2030, with Texas, California, Arizona, New York, and Ohio expected to feel the strain most sharply. That is the concrete shortage behind the policy story: the United States is trying to expand chip production faster than its education, training, and hiring pipelines can produce enough fab technicians, manufacturing specialists, hardware engineers, and process talent.
For WindowsForum readers, the issue is not abstract industrial policy. Chips shape the availability and price of PCs, servers, GPUs, memory, storage, networking gear, and the components inside the systems IT teams buy every year. If new U.S. fabs cannot hire enough people to ramp production on schedule, the result may not be a single dramatic “chip shortage” headline. It may look more like uneven device availability, longer lead times on specific configurations, and less predictable hardware planning.
For the past several years, the U.S. semiconductor revival has been described mostly through capital commitments and factory announcements: Taiwan Semiconductor Manufacturing Co. in Arizona, Micron Technology in New York, Samsung Electronics in Texas, and Intel in Ohio. Those projects still matter. But the Los Angeles Times report, based on analysis involving McKinsey & Co., shifts attention from buildings and subsidies to labor.
Maggie Eastland and Lorelei Smillie, writing for Bloomberg and published by the Los Angeles Times, report that the skilled labor deficit could reach as much as 157,000 full-time workers by 2030. The states expected to be most affected — Texas, California, Arizona, New York, and Ohio — are also central to the new U.S. semiconductor map.
That is the immediate warning. The places gaining fabs are also the places where staffing those fabs could become hardest.
Semiconductor manufacturing is not ordinary factory work with cleaner rooms. A modern fab depends on process control, contamination management, robotics, chemical handling, metrology, equipment maintenance, facilities engineering, yield improvement, and disciplined supply operations. If the workforce is too thin, a fab can be physically complete but still struggle to reach the production levels its owners, customers, and policymakers expect.
Taylor Roundtree, a partner at McKinsey who helped with the analysis, summarized the constraint directly: “There’s just not enough talent to go around.”
That quote matters because it separates the current problem from a simple funding debate. Washington can subsidize fabs. Companies can announce expansions. States can offer land, tax incentives, and infrastructure support. But none of that instantly creates experienced engineers or technicians. The industry needs people who can operate complex facilities, install and maintain tools, tune production, and keep yields moving in the right direction.
According to the Los Angeles Times account, Chips Act funding provided the National Science Foundation with $200 million through 2027 for workforce development, including programs that educate students and train new workers through the National Network for Microelectronics Education. The report says those programs have helped increase the number of technicians available to work at new plants.
That is progress, but it does not erase the central gap. The report says the initiatives have made far less headway on the need for manufacturing and hardware engineers. That distinction matters. Fabs need technicians to operate and maintain production environments, but they also need engineers to solve process, hardware, and manufacturing problems as facilities move from construction to production.
The study cited in the Los Angeles Times report says that, by 2030, about 74% of the semiconductor industry’s unfilled roles will be in manufacturing and 60% will be in engineering. Those categories overlap in practice because semiconductor manufacturing is engineering-heavy. A fab floor is not separate from engineering; it is where engineering discipline is tested continuously.
That is why the gap cannot be fixed by a generic “learn tech” campaign. Semiconductor work requires specialized training, access to equipment, employer involvement, and enough regional depth to support not just one hiring wave but years of operations. Some roles can be taught through short programs. Others require longer academic and practical pathways.
The timing problem is obvious. A fab can be announced in a day, financed over months, and built over several years. A workforce pipeline takes longer. Students need to discover the field, choose relevant classes, enter training programs, gain hands-on experience, and then become productive inside facilities where mistakes can be costly.
These projects should not be treated as interchangeable. TSMC, Micron, Samsung, and Intel have different business models, products, and timelines. But they share a common dependency: none can meet its production goals without enough trained workers.
That is especially important because fabs do not become productive the moment construction ends. They need tools installed, processes tuned, suppliers coordinated, utilities stabilized, workers trained, and yields improved. A labor shortage can slow that transition from impressive facility to reliable production capacity.
The report’s project examples also show why the shortage is regional as well as national. Arizona, New York, Texas, and Ohio are not just names on an investment map. They need local and regional systems that can support hiring over time: community colleges, universities, training providers, employers, suppliers, contractors, and workers willing to live near the projects.
According to the Los Angeles Times report, only about 3% of U.S. engineering students go on to work in the chip industry. Many choose more lucrative software-related fields such as artificial intelligence.
That is one of the most important details in the story. The chip industry is not merely short of workers because students lack technical ability. It is short because students with technical ability often see more attractive opportunities elsewhere. AI and software can offer higher pay, faster career mobility, more geographic flexibility, and a clearer cultural identity for young engineers.
Roundtree’s second quote captures the need for a broader response: “Folks are realizing that the potential gap is so large that they collectively do have to solve it.”
The word “collectively” is doing real work there. A single company can recruit from competitors. It can raise pay. It can sponsor a training program. But if every chipmaker is fighting over the same limited pool, the result is churn rather than a larger workforce. The shortage requires coordination among companies, schools, states, and federal programs.
The industry also has to make a stronger case to students. Semiconductor careers offer technically demanding work with real-world impact, but the path is often less visible than software. Many students understand apps, cloud platforms, games, AI tools, and startup culture. Far fewer have seen a fab, met a process engineer, or understood what a semiconductor manufacturing career looks like.
That kind of outreach can sound symbolic, but visibility is part of the workforce problem. Semiconductor manufacturing has not had a significant U.S. buildout in decades, Roundtree said. If students do not see the industry early, they are less likely to imagine themselves in it later.
This is where the semiconductor sector differs from software. Children and teenagers encounter software constantly. They use apps, games, websites, and AI tools. They hear about programmers, founders, streamers, and cloud platforms. Chip manufacturing is mostly hidden. The most important work happens in controlled facilities that ordinary consumers never enter.
That invisibility affects career choices. A student who has never met a fab technician, hardware engineer, or process engineer may not know the field exists as a practical option. By the time a chipmaker appears at a college career fair, that student may already be committed to AI, cloud software, finance, robotics, aerospace, or consulting.
Early exposure will not close a 157,000-worker gap by itself. But it helps explain why the shortage starts long before companies post job openings. The industry needs students to know that semiconductor work is modern, technical, and connected to the devices and services they use every day.
The report says the authors recommended continued government funding, expanded semiconductor curricula, and earlier exposure to chip-industry careers. Those recommendations point to a pipeline problem, not a one-time hiring problem.
Arizona is a clear example. The report describes TSMC as planning as much as an estimated $265 billion in a dozen chipmaking and packaging facilities in the state. That scale implies more than a single construction project. It implies a regional semiconductor ecosystem that needs workers, suppliers, maintenance capacity, facilities expertise, and managers.
New York faces a similar test through Micron’s $100 billion memory chip production plan. A project at that scale cannot be supported by a small, temporary training push. It requires a durable workforce base that can support production over many years.
Texas already has semiconductor and technology manufacturing experience, but the report still lists it among the states expected to face acute shortages. That is an important reminder: an existing industrial base does not eliminate competition for engineers, technicians, construction labor, and facilities specialists.
Ohio’s case is tied to Intel’s delayed $28 billion investment. A delay can create the impression that there is more time to prepare, but it can also make workforce planning harder if hiring needs arrive after a long period of uncertainty.
California remains a major technology and engineering center, but that does not make it immune. The state has deep talent pools, but also intense competition from AI, software, aerospace, biotech, and other high-paying industries. The shortage there may be less about whether talent exists and more about whether chip manufacturing can attract enough of it.
The semiconductor workforce gap is projected through 2030, so it does not mean every laptop, desktop, server, or component order will suddenly become difficult this year. But it does mean procurement teams should be more careful about assuming that future chip capacity will arrive exactly when factory announcements suggest it will.
The clearest near-term consequence is planning uncertainty. If a company is expecting smoother hardware availability because new U.S. chip projects are underway, this report is a reminder that buildings and production capacity are not the same thing. Labor constraints can slow ramps, and slow ramps can affect the timing and mix of hardware that eventually reaches OEMs and enterprise buyers.
For Windows device planning, that means IT departments should pay closer attention to model availability, component substitutions, and lead times when setting refresh schedules. The risk is not limited to CPUs. PCs and servers also depend on memory, storage controllers, networking components, security chips, and platform support silicon.
For data-center buyers, the report reinforces a broader reality: AI demand is putting pressure on the hardware supply chain at the same time the chip industry is trying to expand domestic production. Even organizations that are not building large AI systems may feel secondary effects through server availability, cloud capacity decisions, or pricing for higher-end hardware.
For procurement teams, the practical move is to keep options open. Organizations that rely on a single approved model, a narrow configuration list, or a tightly timed refresh window may have less flexibility if specific components or systems become harder to source. The report does not prove that a specific Windows device class will be constrained, so the right response is measured preparation rather than overreaction.
The report says Chips Act-funded programs have helped increase the number of technicians available to work at new plants. That is meaningful. But the same report says the need for manufacturing and hardware engineers remains a major issue. That split should shape the public discussion.
Not all semiconductor labor gaps are the same. Construction workers build fabs. Technicians help operate and maintain them. Engineers solve process, hardware, and manufacturing problems. Each group requires a different pipeline, and success in one category does not automatically solve the others.
The 157,000-worker figure is therefore not just a staffing estimate. It is a test of whether the United States can rebuild the human infrastructure behind advanced manufacturing. Fabs require more than capital expenditure. They require schools, instructors, lab access, internships, employer commitments, regional coordination, and students who see semiconductor work as a serious career.
The compensation and career proposition also matter. If only about 3% of U.S. engineering students enter the chip industry while many choose more lucrative software-related fields such as AI, the industry has to compete more effectively for talent. That does not mean every semiconductor job can match the hottest AI role. It does mean chipmakers must make the path more visible, more attractive, and more credible to students weighing their options.
This is the gap between industrial policy as announced and industrial policy as delivered. The United States can fund factories, but the market only feels the effect when those factories can produce reliably at scale.
For WindowsForum readers, the takeaway is practical: treat the chip workforce shortage as a medium-term hardware planning risk. Do not assume that every announced fab will translate into predictable device and server availability on the timeline buyers would prefer. Build flexibility into refresh schedules, keep approved hardware alternatives ready, and ask vendors direct questions about lead times and configuration risk.
The U.S. chip buildout is still moving forward. But the report makes clear that the next bottleneck may be less visible than a missing factory or a delayed subsidy. It may be the engineer who was never trained, the technician who chose another field, or the student who never learned that semiconductor work was an option at all.
For WindowsForum readers, the issue is not abstract industrial policy. Chips shape the availability and price of PCs, servers, GPUs, memory, storage, networking gear, and the components inside the systems IT teams buy every year. If new U.S. fabs cannot hire enough people to ramp production on schedule, the result may not be a single dramatic “chip shortage” headline. It may look more like uneven device availability, longer lead times on specific configurations, and less predictable hardware planning.
The Chip Boom Has Found Its Bottleneck, and It Is Human
For the past several years, the U.S. semiconductor revival has been described mostly through capital commitments and factory announcements: Taiwan Semiconductor Manufacturing Co. in Arizona, Micron Technology in New York, Samsung Electronics in Texas, and Intel in Ohio. Those projects still matter. But the Los Angeles Times report, based on analysis involving McKinsey & Co., shifts attention from buildings and subsidies to labor.Maggie Eastland and Lorelei Smillie, writing for Bloomberg and published by the Los Angeles Times, report that the skilled labor deficit could reach as much as 157,000 full-time workers by 2030. The states expected to be most affected — Texas, California, Arizona, New York, and Ohio — are also central to the new U.S. semiconductor map.
That is the immediate warning. The places gaining fabs are also the places where staffing those fabs could become hardest.
Semiconductor manufacturing is not ordinary factory work with cleaner rooms. A modern fab depends on process control, contamination management, robotics, chemical handling, metrology, equipment maintenance, facilities engineering, yield improvement, and disciplined supply operations. If the workforce is too thin, a fab can be physically complete but still struggle to reach the production levels its owners, customers, and policymakers expect.
Taylor Roundtree, a partner at McKinsey who helped with the analysis, summarized the constraint directly: “There’s just not enough talent to go around.”
That quote matters because it separates the current problem from a simple funding debate. Washington can subsidize fabs. Companies can announce expansions. States can offer land, tax incentives, and infrastructure support. But none of that instantly creates experienced engineers or technicians. The industry needs people who can operate complex facilities, install and maintain tools, tune production, and keep yields moving in the right direction.
Washington Bought Fabs Faster Than Schools Can Produce Fab Workers
The 2022 Chips and Science Act was designed to strengthen U.S. semiconductor manufacturing and the ecosystem around it. The visible side of that effort is the factory buildout. The slower side is workforce development.According to the Los Angeles Times account, Chips Act funding provided the National Science Foundation with $200 million through 2027 for workforce development, including programs that educate students and train new workers through the National Network for Microelectronics Education. The report says those programs have helped increase the number of technicians available to work at new plants.
That is progress, but it does not erase the central gap. The report says the initiatives have made far less headway on the need for manufacturing and hardware engineers. That distinction matters. Fabs need technicians to operate and maintain production environments, but they also need engineers to solve process, hardware, and manufacturing problems as facilities move from construction to production.
The study cited in the Los Angeles Times report says that, by 2030, about 74% of the semiconductor industry’s unfilled roles will be in manufacturing and 60% will be in engineering. Those categories overlap in practice because semiconductor manufacturing is engineering-heavy. A fab floor is not separate from engineering; it is where engineering discipline is tested continuously.
That is why the gap cannot be fixed by a generic “learn tech” campaign. Semiconductor work requires specialized training, access to equipment, employer involvement, and enough regional depth to support not just one hiring wave but years of operations. Some roles can be taught through short programs. Others require longer academic and practical pathways.
The timing problem is obvious. A fab can be announced in a day, financed over months, and built over several years. A workforce pipeline takes longer. Students need to discover the field, choose relevant classes, enter training programs, gain hands-on experience, and then become productive inside facilities where mistakes can be costly.
The Biggest Projects Are Also the Biggest Labor Tests
The Los Angeles Times report points to several large U.S. semiconductor projects that illustrate the scale of the challenge. The numbers are large, but the labor question is the same in each case: can the region and the industry provide enough qualified people when production needs to ramp?| Company | Planned U.S. project | Location | Stated scale in the report | Production focus stated in the report | Workforce risk implied by the report |
|---|---|---|---|---|---|
| Taiwan Semiconductor Manufacturing Co. | Chipmaking and packaging facilities | Arizona | As much as an estimated $265 billion in a dozen facilities | Chipmaking and packaging | Arizona is among the states expected to face acute shortages |
| Micron Technology Inc. | Memory chip production plan | New York | $100 billion | Memory chips | New York is among the states expected to face acute shortages |
| Samsung Electronics Co. | Logic-chip facility | Texas | Not specified in the provided source material | Logic chips | Texas is among the states expected to face acute shortages |
| Intel Corp. | Delayed investment | Ohio | $28 billion | Not specified in the provided source material | Shortages are expected once production ramps |
That is especially important because fabs do not become productive the moment construction ends. They need tools installed, processes tuned, suppliers coordinated, utilities stabilized, workers trained, and yields improved. A labor shortage can slow that transition from impressive facility to reliable production capacity.
The report’s project examples also show why the shortage is regional as well as national. Arizona, New York, Texas, and Ohio are not just names on an investment map. They need local and regional systems that can support hiring over time: community colleges, universities, training providers, employers, suppliers, contractors, and workers willing to live near the projects.
AI Is Pulling Talent Away From the Industry That AI Needs
The semiconductor labor shortage is tied to AI in a direct way. AI systems require chips, memory, servers, networking equipment, and data-center infrastructure. But the same AI boom that increases demand for semiconductors also competes for engineering talent.According to the Los Angeles Times report, only about 3% of U.S. engineering students go on to work in the chip industry. Many choose more lucrative software-related fields such as artificial intelligence.
That is one of the most important details in the story. The chip industry is not merely short of workers because students lack technical ability. It is short because students with technical ability often see more attractive opportunities elsewhere. AI and software can offer higher pay, faster career mobility, more geographic flexibility, and a clearer cultural identity for young engineers.
Roundtree’s second quote captures the need for a broader response: “Folks are realizing that the potential gap is so large that they collectively do have to solve it.”
The word “collectively” is doing real work there. A single company can recruit from competitors. It can raise pay. It can sponsor a training program. But if every chipmaker is fighting over the same limited pool, the result is churn rather than a larger workforce. The shortage requires coordination among companies, schools, states, and federal programs.
The industry also has to make a stronger case to students. Semiconductor careers offer technically demanding work with real-world impact, but the path is often less visible than software. Many students understand apps, cloud platforms, games, AI tools, and startup culture. Far fewer have seen a fab, met a process engineer, or understood what a semiconductor manufacturing career looks like.
The Semiconductor Career Pipeline Broke Quietly
One of the more concrete examples in the Los Angeles Times story is an Arizona outreach effort where elementary school students handled semiconductor equipment and tried on a white bunny suit, the protective coverall used in fab environments to reduce contamination risk.That kind of outreach can sound symbolic, but visibility is part of the workforce problem. Semiconductor manufacturing has not had a significant U.S. buildout in decades, Roundtree said. If students do not see the industry early, they are less likely to imagine themselves in it later.
This is where the semiconductor sector differs from software. Children and teenagers encounter software constantly. They use apps, games, websites, and AI tools. They hear about programmers, founders, streamers, and cloud platforms. Chip manufacturing is mostly hidden. The most important work happens in controlled facilities that ordinary consumers never enter.
That invisibility affects career choices. A student who has never met a fab technician, hardware engineer, or process engineer may not know the field exists as a practical option. By the time a chipmaker appears at a college career fair, that student may already be committed to AI, cloud software, finance, robotics, aerospace, or consulting.
Early exposure will not close a 157,000-worker gap by itself. But it helps explain why the shortage starts long before companies post job openings. The industry needs students to know that semiconductor work is modern, technical, and connected to the devices and services they use every day.
The report says the authors recommended continued government funding, expanded semiconductor curricula, and earlier exposure to chip-industry careers. Those recommendations point to a pipeline problem, not a one-time hiring problem.
The States Winning Fab Projects May Inherit Fab Strain
Texas, California, Arizona, New York, and Ohio are expected to be among the states where the shortage is most acute. That matters because states have competed aggressively for semiconductor investment. Winning a fab project is only the first phase. The next phase is proving that the region can absorb it.Arizona is a clear example. The report describes TSMC as planning as much as an estimated $265 billion in a dozen chipmaking and packaging facilities in the state. That scale implies more than a single construction project. It implies a regional semiconductor ecosystem that needs workers, suppliers, maintenance capacity, facilities expertise, and managers.
New York faces a similar test through Micron’s $100 billion memory chip production plan. A project at that scale cannot be supported by a small, temporary training push. It requires a durable workforce base that can support production over many years.
Texas already has semiconductor and technology manufacturing experience, but the report still lists it among the states expected to face acute shortages. That is an important reminder: an existing industrial base does not eliminate competition for engineers, technicians, construction labor, and facilities specialists.
Ohio’s case is tied to Intel’s delayed $28 billion investment. A delay can create the impression that there is more time to prepare, but it can also make workforce planning harder if hiring needs arrive after a long period of uncertainty.
California remains a major technology and engineering center, but that does not make it immune. The state has deep talent pools, but also intense competition from AI, software, aerospace, biotech, and other high-paying industries. The shortage there may be less about whether talent exists and more about whether chip manufacturing can attract enough of it.
What It Means This Year for WindowsForum Readers
For IT teams, this report should be read as a hardware-roadmap warning, not as a reason to panic-buy.The semiconductor workforce gap is projected through 2030, so it does not mean every laptop, desktop, server, or component order will suddenly become difficult this year. But it does mean procurement teams should be more careful about assuming that future chip capacity will arrive exactly when factory announcements suggest it will.
The clearest near-term consequence is planning uncertainty. If a company is expecting smoother hardware availability because new U.S. chip projects are underway, this report is a reminder that buildings and production capacity are not the same thing. Labor constraints can slow ramps, and slow ramps can affect the timing and mix of hardware that eventually reaches OEMs and enterprise buyers.
For Windows device planning, that means IT departments should pay closer attention to model availability, component substitutions, and lead times when setting refresh schedules. The risk is not limited to CPUs. PCs and servers also depend on memory, storage controllers, networking components, security chips, and platform support silicon.
For data-center buyers, the report reinforces a broader reality: AI demand is putting pressure on the hardware supply chain at the same time the chip industry is trying to expand domestic production. Even organizations that are not building large AI systems may feel secondary effects through server availability, cloud capacity decisions, or pricing for higher-end hardware.
For procurement teams, the practical move is to keep options open. Organizations that rely on a single approved model, a narrow configuration list, or a tightly timed refresh window may have less flexibility if specific components or systems become harder to source. The report does not prove that a specific Windows device class will be constrained, so the right response is measured preparation rather than overreaction.
Action Checklist for Admins and Procurement Teams
- Review refresh schedules for single-window risk. If a large Windows hardware refresh depends on one ordering window, identify whether the schedule can be staggered without disrupting support or security requirements.
- Ask OEMs and resellers about lead-time sensitivity. Focus on concrete questions: which approved models are easiest to source, which configurations have longer lead times, and which component choices create the most substitution risk.
- Preapprove practical alternatives. Maintain a short list of acceptable substitute Windows endpoint and server configurations so procurement is not forced to make policy decisions during a shortage or delay.
- Separate routine refresh demand from AI-related purchases. AI PCs, GPU servers, memory-heavy systems, and standard office endpoints may face different supply pressures. Tracking them separately gives finance and IT a clearer view of demand.
- Avoid treating factory announcements as guaranteed capacity. When planning multi-year hardware budgets, remember that fabs must still hire, ramp, and stabilize production before new capacity affects the market.
The Report Is a Warning Against Subsidy Theater
The easiest political response to the chip-worker shortage is to point to existing subsidies and training programs. The harder question is whether those programs are producing enough workers in the right roles quickly enough.The report says Chips Act-funded programs have helped increase the number of technicians available to work at new plants. That is meaningful. But the same report says the need for manufacturing and hardware engineers remains a major issue. That split should shape the public discussion.
Not all semiconductor labor gaps are the same. Construction workers build fabs. Technicians help operate and maintain them. Engineers solve process, hardware, and manufacturing problems. Each group requires a different pipeline, and success in one category does not automatically solve the others.
The 157,000-worker figure is therefore not just a staffing estimate. It is a test of whether the United States can rebuild the human infrastructure behind advanced manufacturing. Fabs require more than capital expenditure. They require schools, instructors, lab access, internships, employer commitments, regional coordination, and students who see semiconductor work as a serious career.
The compensation and career proposition also matter. If only about 3% of U.S. engineering students enter the chip industry while many choose more lucrative software-related fields such as AI, the industry has to compete more effectively for talent. That does not mean every semiconductor job can match the hottest AI role. It does mean chipmakers must make the path more visible, more attractive, and more credible to students weighing their options.
Timeline: From Factory Announcement to Workforce Reality
| Stage | What usually gets attention | Workforce issue that can be overlooked |
|---|---|---|
| Subsidy or investment announcement | Dollar amount, location, political support | Whether the region has enough relevant training capacity |
| Construction | Site work, buildings, utilities, equipment plans | Competition for construction and facilities-related labor |
| Tool installation and ramp preparation | Production targets and customer expectations | Need for technicians, engineers, vendors, and experienced managers |
| Early production | First output and public milestones | Yield learning, process stability, maintenance, and staffing depth |
| High-volume production | Capacity claims and market impact | Whether the workforce can sustain operations over time |
The Bottom Line
The Los Angeles Times report is a reminder that the semiconductor race is not only about money, land, equipment, or geopolitics. It is also about whether enough people can be trained, recruited, and retained to make new capacity real.For WindowsForum readers, the takeaway is practical: treat the chip workforce shortage as a medium-term hardware planning risk. Do not assume that every announced fab will translate into predictable device and server availability on the timeline buyers would prefer. Build flexibility into refresh schedules, keep approved hardware alternatives ready, and ask vendors direct questions about lead times and configuration risk.
The U.S. chip buildout is still moving forward. But the report makes clear that the next bottleneck may be less visible than a missing factory or a delayed subsidy. It may be the engineer who was never trained, the technician who chose another field, or the student who never learned that semiconductor work was an option at all.
References
- Primary source: Los Angeles Times
Published: Wed, 08 Jul 2026 13:58:57 GMT
Loading…
www.latimes.com - Related coverage: tomshardware.com
Loading…
www.tomshardware.com - Related coverage: tradingkey.com
Loading…
www.tradingkey.com - Related coverage: investing.com
Loading…
www.investing.com - Related coverage: abq.org
Loading…
www.abq.org - Related coverage: techcrunch.com
Loading…
techcrunch.com