Friday Founders: Issue #8 | Highlights from 8–14 August 2025

Hi Founders,

It’s Friday, 15 August 2025, and the tape didn’t slow down: hedge funds piled back into Big Tech, crypto hit a fresh high, and AI infra kept gobbling capex. Here are 10 stories to scan before the weekend.

1) 📈 Hedge funds double down on Big Tech amid the AI boom

TL; DR: New 13F analysis shows hedge funds rotating back into the largest U.S. tech names as AI spending keeps growth and margins elevated. The re-risking follows hotter-than-expected July PPI data that briefly rattled markets before money flowed to perceived AI winners. Expect more factor crowding in megacap tech through Q3.

Why you should care: If you sell to Fortune 500 tech or build on their platforms, budget cycles may stay friendly even if broader conditions wobble. Startup valuations that benchmark to public comps could see a tailwind while this rotation holds. Plan launches and fundraising against that window.

Read more → Reuters 

2) 🛡️ National-security scrutiny: the “Nvidia deal” alarms lawmakers

TL; DR: An unusual arrangement involving Nvidia and the U.S. administration triggered bipartisan concern over corporate governance and national-security risk. Analysts warn that any perception of preferential treatment could draw tighter controls on AI hardware supply chains. Export-control and verification proposals are back on the table.

Why you should care: If your GTM depends on training/inference in sensitive regions, expect more paperwork and geo-fencing. Build “where did this run?” attestations into your infra story now. Keep a contingency plan for supply or policy shocks.

Read more → Reuters

3) ⚖️ EU AI Act: GPAI obligations are live, with fresh guidance

TL; DR: As of August 2, new obligations for general-purpose AI models apply in the EU; pre-existing models have a longer runway, but enforcement structures are in place. The Commission published guidelines clarifying scope and documentation expectations, and firms are being told to prepare audits. Law firms are flagging national authorities and reporting duties that already kicked in.

Why you should care: Selling into the EU now requires training-data summaries, evaluation methods, and red-team logs plus a plan if you’re deemed “systemic risk.” Downstream startups should demand attestations from model vendors. Budget time for model-card updates as templates evolve.

Read more → DLA Piper

4) 🩺 Abridge (healthcare AI, $5.3B) eyes acquisitions

TL; DR: Clinical-documentation startup Abridge says it will spend the bulk of its fresh cash on core tech and allocate ~20% to M&A, targeting teams with talent or IP in notes, billing, and decision support. Competition is intensifying against incumbents like Epic and fast-moving startups in ambient scribing. Headcount and model-training plans are ramping.

Why you should care: If you’re building in clinical NLP, revenue cycle, or care ops, Abridge could be a strategic acquirer or a formidable rival. Strengthen your differentiation (data rights, specialty depth, integrations) before conversations. Hospitals want fewer vendors; partner-or-be-bought dynamics are rising.

Read more → Business Insider

5) 🏀 PE money buys the Portland Trail Blazers (>$4B)

TL; DR: Tom Dundon struck a deal to purchase the NBA’s Portland Trail Blazers from the Paul Allen estate for north of $4B, with additional backers lining up. The team no longer owns its arena, which alters the usual real-estate economics. The buyer pledged to keep the team in Portland, but stadium funding debates loom.

Why you should care: Big private-market checks are still clearing useful context for late-stage founders eyeing minority PE rounds or liquidity options. Sports M&A often precedes broader risk-on sentiment. Expect more crossover investors to peek at growth-stage software again.

Read more → Axios

6) 🏭 Foxconn’s profit jumps as server revenue surpasses smartphones

TL; DR: Foxconn reported a 27% profit boost, crediting AI server demand that now outpaces traditional smartphone assembly revenue. The result came during a choppy macro session triggered by hotter U.S. producer-price data. AI infrastructure continues to be the manufacturing bright spot.

Why you should care: Hardware founders: lead times and pricing for racks, power, and thermals will stay tight. If your product relies on data-center deployments, align with integrators early. For mobile-first startups, note where talent and capex are flowing.

Read more → Reuters

7) ₿ Crypto rips: Bitcoin > $124,000; Bullish pops on NYSE debut

TL; DR: Bitcoin notched a new all-time high above $124k while exchange Bullish surged in its public debut investors cited improving U.S. regulatory outlook plus rate-cut hopes. The moves came despite macro jitters in equities. Liquidity and retail attention are back in force.

Why you should care: If you’re building on crypto rails (payments, loyalty, remittance), rising prices and a friendlier tone can speed pilots. But volatility can whiplash CAC and fraud bake that into models. Keep compliance checklists current as rules harden.

Read more → Reuters

8) 🧠 RISC-V/AI chipmaker Rivos lines up a $500M round (in market)

TL; DR: Axios reports Rivos is targeting roughly $500M, adding fuel to the alternative-chip race against x86/ARM incumbents. The company pitches custom silicon for on-prem and edge AI. The raise comes amid heavy sovereign and strategic interest in compute independence.

Why you should care: Builders of inference-heavy products should model a multi-ISA future. Portability across NVIDIA, AMD, and emerging RISC-V options will matter to enterprise buyers. Keep your kernels, compilers, and quantization paths versatile.

Read more → The Information

9) 🏥 Early-stage med-AI: Better Medicine raises pre-seed

TL; DR: Estonia’s Better Medicine raised €1M to expand its AI kidney-cancer detection work beyond the Baltics and pursue U.S. partnerships. The team is pushing toward clinical validation and insurer engagement. European health-AI continues to seed new vertical specialists.

Why you should care: Even small checks signal where regulators and payers might be more receptive. If you’re a platform, these ultra-focused teams can be high-leverage partners. For founders in imaging, nail data provenance and bias controls early.

Read more → EU-Startups

10) 🧪 Open models + science: NVIDIA & U.S. NSF back Ai2 effort

TL;DR: NVIDIA’s newsroom highlighted a collaboration with the National Science Foundation to support the Allen Institute for AI (Ai2) in developing open AI models for U.S. scientific leadership. Fellowships and grants are rolling in alongside compute support. The push underscores open-science momentum alongside proprietary stacks.

Why you should care: If you work in scientific or technical AI, expect more high-quality open models and data to plug into. That can shrink your time-to-prototype, but also raises the bar for defensible moats. Differentiate with domain data, workflows, and validation not just model choice.

Read more → NVIDIA Newsroom