In partnership with
Good morning.
Nvidia just acquired AI chip challenger Groq for $20 billion, removing one of its fastest-growing competitors. The U.S. accused China of unfair chip trade practices as Trump and Xi navigate a fragile truce. And the Trump administration's H-1B visa fee increase to $100,000 just got judge backing.
Meanwhile, Starlink hit 9 million active customers worldwide as growth accelerates. And Bitcoin miners are pivoting to AI infrastructure, lifting stocks even as crypto prices lag.
This week showed Nvidia consolidating power, immigration policy reshaping tech talent, and infrastructure providers (Starlink, Bitcoin miners) finding new revenue streams.
Your Sunday rundown:
Nvidia acquires Groq for $20B (removes challenger)
U.S. accuses China of unfair chip trade
H-1B visa fees jump to $100K (judge approved)
Starlink hits 9M customers (accelerating growth)
Bitcoin miners pivot to AI (new revenue model)
🎁 Before you keep reading, don't miss this special opportunity from Westbound & Down
Colorado’s Most-Awarded Brewery Did Something Totally Unique
Some companies make lofty promises to investors and never deliver. Others use those dollars to unlock new levels of scale.
That’s Westbound & Down’s story. Already Colorado’s most-awarded craft brewery, they opened their doors to investors for the first time to help open a flagship Denver-metro-area location.
With 2,800% distribution growth since 2019 and a retail partnership with Whole Foods, it’s no shock investors maxed out that campaign in less than 60 days.
But it’s what comes next that’s even more exciting. Fresh off Brewery of the Year honors at the 2025 Great American Beer Festival, W&D is scaling toward 4X distribution growth by 2028.
This is a paid advertisement for Westbound & Down’s Regulation CF Offering. Please read the offering circular at https://invest.westboundanddown.com/
Hey there, Alberto here.
Every Sunday I break down the biggest moves in tech and startups—the M&A that reshapes markets, the policy changes that affect talent, and the pivots that create new opportunities.
This week had major consolidation (Nvidia-Groq), escalating trade tensions, and infrastructure companies finding new markets.
Let's dive in.
1. Nvidia Acquires Groq for $20B; Removes Fastest Challenger
Nvidia just acquired Groq, one of its most promising AI chip competitors, for $20 billion.
Groq had developed chips specifically optimized for AI inference (running trained models), claiming 10X faster performance than Nvidia's GPUs for certain workloads.
The acquisition removes Nvidia's fastest-growing challenger and consolidates its dominance in AI infrastructure.
Why this matters: This isn't just an acquisition. It's Nvidia preventing competition before it scales.
Groq was gaining traction with AI companies looking for alternatives to Nvidia's expensive H100 and H200 GPUs. Companies like Meta, Anthropic, and OpenAI were testing Groq chips for inference workloads.
If Groq had scaled independently, it could have captured 10-20% of the inference market—worth tens of billions annually.
By acquiring Groq for $20B, Nvidia:
Removes a credible competitor
Acquires Groq's technology and talent
Controls both training (H100/H200) and inference (Groq) markets
Maintains pricing power across the entire AI stack
The competitive landscape:
With Groq gone, Nvidia's remaining chip challengers are:
AMD (MI300X chips, gaining share but still small)
Google (TPUs, internal use only)
Amazon (Trainium/Inferentia, internal use only)
Cerebras (high-end research, niche market)
None have Groq's momentum or commercial traction. Nvidia just cleared the board.
Investment angle: When dominant platforms acquire their fastest-growing competitors, it signals they see real threats. Nvidia wouldn't pay $20B for Groq if it wasn't worried. This consolidation is good for Nvidia shareholders (removes competition) but bad for AI companies seeking alternatives (fewer choices, higher prices). For investors: Nvidia's moat just got wider. For AI startups: compute costs aren't coming down anytime soon.
2. U.S. Accuses China of Unfair Chip Trade Amid Trump-Xi Truce
The U.S. government accused China of unfair trade practices in the semiconductor industry, even as Trump and Xi attempt to maintain a fragile trade truce.
The accusations focus on:
Chinese subsidies to domestic chip makers (SMIC, others)
Forced technology transfers from foreign companies
Below-cost pricing to gain market share
This comes as the Trump administration weighs additional export restrictions on advanced chip-making equipment to China.
Why this matters: The chip war between the U.S. and China is escalating, not easing.
Despite Trump and Xi's public truce, the underlying competition for semiconductor dominance continues. Here's why chips are geopolitical:
1. AI requires chips. Whoever controls advanced chips controls AI development. China doesn't want to depend on Nvidia (U.S.) for AI infrastructure.
2. Military applications. Advanced chips power autonomous weapons, surveillance systems, and cyber warfare. National security depends on domestic chip production.
3. Economic leverage. The country that manufactures chips has leverage over everyone else. China manufactures 70% of low-end chips. The U.S. designs most advanced chips (via Nvidia, AMD, Intel).
The fragile equilibrium:
Trump wants trade with China (helps U.S. economy). But he also wants to limit China's access to advanced chips (national security). Those goals conflict.
The U.S. accusation signals that even during a "truce," competition for chip dominance continues.
Investment angle: Semiconductor companies are caught in geopolitical crossfire. U.S. chip makers (Nvidia, AMD, Intel) face potential loss of Chinese market. Chinese chip makers (SMIC) face technology restrictions. Taiwan (TSMC) is the critical bottleneck—whoever controls Taiwan controls advanced chip production. For investors: Semiconductor supply chains are fragmenting by geography. Companies with diversified manufacturing (Intel building U.S. fabs, TSMC building Arizona fabs) are positioned better than those dependent on single regions.
3. H-1B Visa Fees Jump to $100K—Judge Backs Trump Policy
The Trump administration's increase of H-1B visa fees to $100,000 per application just received backing from a federal judge, clearing the way for implementation.
Previously, H-1B visa fees were $5,000-$10,000 per application. The new $100K fee is a 10-20X increase.
The policy aims to discourage companies from hiring foreign workers and encourage hiring U.S. citizens instead.
Why this matters: This fundamentally changes the economics of hiring international talent in tech.
The math before:
Company hires engineer from India/China on H-1B visa. Pays $10K in fees, $120K salary. Total cost: $130K/year.
The math after:
Same engineer now costs $100K in fees (one-time), $120K salary. First-year cost: $220K. That's 69% more expensive.
Who this hurts:
Startups: Can't afford $100K visa fees when hiring is tight. Will struggle to recruit international talent.
Big Tech: Can absorb costs but will favor U.S. citizens to avoid fees. International hiring will decline.
International talent: H-1B visas become less accessible. Skilled workers may choose Canada, UK, or EU instead of U.S.
Who this helps:
U.S. tech workers: Less competition from international talent should increase wages.
Immigration lawyers: More complexity = more legal fees.
Canada/UK/EU: Countries with friendlier immigration policies will attract talent the U.S. rejects.
The unintended consequences:
Many of the most successful tech companies were founded or co-founded by immigrants (Google, Tesla, Nvidia, Zoom, WhatsApp). Making immigration harder could reduce the next generation of founders choosing the U.S.
Investment angle: Startups will struggle more to recruit technical talent, especially AI/ML engineers (where international talent is heavily concentrated). Companies with offshore engineering teams (India, Eastern Europe) may become more attractive. Canada's tech ecosystem could see significant growth as talent diverts there. For U.S. tech workers, this should lift wages in the short term—but may slow innovation long term if the best talent goes elsewhere.
4.
Starlink Hits 9M Active Customers: Accelerating Growth
Starlink just hit 9 million active customers worldwide, up from 7 million in Q3 2025.
That's 2 million net adds in one quarter—the fastest growth rate in the company's history.
Revenue is estimated at $9-10 billion annually ($100/month average per customer × 9M customers × 12 months).
Why this matters: Starlink is becoming a real business, not just an Elon side project.
The growth drivers:
1. Underserved markets: Rural areas, maritime, aviation. Places where traditional internet doesn't reach or is prohibitively expensive.
2. Natural disasters: When hurricanes, wildfires, or earthquakes take out ground infrastructure, Starlink remains operational. Governments and emergency services are buying.
3. Military/government: U.S. military, NATO, and allies use Starlink for communications in remote areas and conflict zones (Ukraine proved its effectiveness).
4. Developing countries: Africa, Southeast Asia, Latin America—regions with poor internet infrastructure are leapfrogging to satellite.
The path to profitability:
At 9M customers × $100/month = $900M/month = $10.8B/year in revenue.
Starlink's costs:
Satellite launches (declining as SpaceX reuses rockets)
Ground stations (one-time infrastructure)
Customer terminals (subsidized initially, now ~breakeven)
SpaceX CEO Elon Musk has said Starlink will be profitable in 2025. At $10B+ revenue with improving margins, that's credible.
Investment angle: Starlink is becoming SpaceX's cash cow. It generates predictable, recurring revenue (subscriptions) while rocket launches are lumpier. If SpaceX IPOs (targeting 2026 at $1.5T valuation), Starlink will be a major valuation driver. Comparable: Viasat (satellite internet competitor) trades at ~2X revenue. At $10B revenue, Starlink alone could be worth $20-30B. That's just one piece of SpaceX's $1.5T target valuation. For investors: Starlink validates that LEO satellite internet works at scale. Companies building adjacent infrastructure (ground stations, terminals, competing constellations like Amazon's Kuiper) should see increased investor interest.
5. Bitcoin Miners Pivot to AI—Lifting Stocks as Crypto Lags
Bitcoin miners are pivoting to AI infrastructure, converting mining facilities into data centers for AI compute.
The strategy: Use existing infrastructure (buildings, power, cooling) and swap Bitcoin mining rigs for AI GPUs (Nvidia H100s, H200s).
Result: Mining company stocks are up 30-50% this quarter even as Bitcoin prices lag.
Why this matters: This is one of the smartest pivots in recent tech history.
The economics:
Bitcoin mining (2025):
Highly competitive (difficulty increased 10X in 5 years)
Thin margins (electricity costs eat most revenue)
Volatile (Bitcoin price swings 30-50% quarterly)
Uncertain regulatory environment (some states banning mining)
AI compute (2025):
High demand (AI companies desperate for GPU capacity)
Better margins (charge $2-3/hour per H100 GPU vs pennies per Bitcoin hash)
Predictable revenue (long-term contracts with AI companies)
Favorable regulation (governments want domestic AI infrastructure)
The infrastructure advantage:
Bitcoin miners already have:
Cheap power contracts (often hydro, nuclear, or stranded energy)
Cooling systems (critical for GPUs)
Buildings and security
Electrical infrastructure
Converting to AI data centers requires swapping hardware, not rebuilding from scratch.
Who's pivoting:
Core Scientific: Signed deals with CoreWeave (AI cloud) to host 200MW of GPU capacity.
Riot Platforms: Converting Texas facilities to AI/HPC infrastructure.
Marathon Digital: Exploring dual-use (Bitcoin + AI) facilities.
The trend: As Bitcoin mining becomes less profitable, miners are becoming AI infrastructure providers. They're essentially landlords renting compute to AI companies.
Investment angle: This validates that AI compute demand is real and supply-constrained. If Bitcoin miners can profitably pivot to AI infrastructure, it means AI companies are desperate enough for capacity to sign deals with non-traditional providers. For investors: Bitcoin mining stocks are re-rating from "crypto plays" to "AI infrastructure plays." That's a higher valuation multiple (AI infrastructure trades at 10-15X revenue, Bitcoin miners traded at 2-3X). Watch for more miners announcing AI pivots—those stocks will see immediate re-ratings.
What This Week Tells Us
Five stories, one theme: Consolidation, conflict, and creative pivots.
Nvidia consolidated by acquiring Groq ($20B). The U.S. and China are escalating chip trade conflict. H-1B visa fees hit $100K, reshaping talent competition. Starlink hit 9M customers, proving satellite internet scales. And Bitcoin miners pivoted to AI, finding a better business model.
Key takeaways:
Nvidia tightens grip - Groq acquisition removes fastest challenger, controls training + inference
Chip wars escalating - U.S.-China conflict continuing despite Trump-Xi truce
Immigration reshaping tech talent - $100K H-1B fees will slow international hiring, may push talent to Canada/EU
Starlink validates satellite internet - 9M customers, $10B revenue, path to profitability clear
Bitcoin miners find new life - Pivoting to AI infrastructure, stocks up 30-50% as they re-rate
The winners: Companies with moats (Nvidia), alternative infrastructure (Starlink), and those pivoting to growth markets (Bitcoin miners).
The losers: AI companies facing higher compute costs, startups unable to hire international talent, and countries losing tech talent to immigration restrictions.
Founderscrowd Premium:
$40/month (beta pricing) before we increase to $120/month.
✅ Weekly pre-IPO opportunities
✅ Investment banker-quality memos
✅ Direct team access for Q&A
✅ Community of smart investors
✅ Lock in beta pricing forever
Enjoy the rest of your Sunday.
See you Tuesday for the year-end wrap-up.
Stay sharp,
Alberto Rosado