Chip and tech stocks rally in AI-trade revival

On Monday, July 6, 2026, the AI trade found its footing on the back of two concrete corporate events rather than vague optimism. Broadcom and Apple agreed to extend their partnership through 2031, locking in the supply of custom radio-frequency, Wi-Fi, Bluetooth and networking chips across Apple's product lineup ^[1]. Because Apple accounts for roughly a fifth of Broadcom's revenue, the reassurance of durable demand mattered well beyond a single contract, and the stock rose about 4% to $375.40 in early trading ^[2].

The second catalyst came from the memory side of the industry. SK Hynix kicked off formal marketing for a roughly $28 billion US listing, selling American depositary receipts under the Nasdaq ticker SKHY and drawing about $7 billion in investor interest before shares even started trading ^[3]. An offering that rivals Saudi Aramco's 2019 IPO in size is not a defensive move - it is a bet that appetite for the high-bandwidth memory feeding AI data centers is nowhere near exhausted, with proceeds earmarked for new fabs and ASML EUV equipment ^[4].

The most important shift in this rally is not that chip stocks went up, but which chip stocks went up. Through the second quarter of 2026, non-Nvidia chipmakers - Micron, Intel and AMD - added roughly $2 trillion in combined market value as investors widened their AI portfolios beyond the single name that defined the trade's first phase ^[5]. Micron is up more than 300% year to date and the Philadelphia Semiconductor Index has climbed about 47%, evidence that the money is flowing into the whole supply chain rather than one marquee GPU maker ^[6].

Wall Street's forecasters have leaned into the broadening. Bank of America hiked its 2026 global chip-sales forecast to $1.3 trillion, roughly 30% growth, and framed the boom as being led jointly by Nvidia and Broadcom rather than any one company ^[7]. That reframing matters: a trade that rests on two or three pillars is harder to topple with a single disappointing earnings print than one that lives and dies with Nvidia alone.

The revival only makes sense against the crash that preceded it. In early June 2026, a single session erased roughly $1.4 trillion in AI-chip value - the sector's worst day since 2020 - after Broadcom merely reiterated rather than raised its full-year guidance ^[8]. That air-pocket reset expectations low enough that Monday's dip-buyers had room to act without chasing all-time highs.

When the rebound came, it was sharp: chipmakers such as Nvidia and Micron climbed more than 5% in the recovery session as easing Iran-Israel tensions and continued hyperscaler capex gave risk appetite cover to return ^[6]. The sequence is the point. A market that fell hard on a guidance non-event, then bought the dip weeks later on two supply-side deals, is one still trading on conviction about AI demand rather than fresh fundamentals - and the next concrete test arrives quickly, with SK Hynix's ADRs set to begin trading on the Nasdaq under the ticker SKHY on July 10 ^[3].

Not everyone reads the bounce as a green light. The dip-buyers have Wall Street cover - JPMorgan strategists were urging clients to buy the pullback in chip stocks, a call amplified widely across finance-focused X accounts - but the caution is just as visible. Morgan Stanley strategist Mike Wilson warned that markets rarely move in a straight line at the pace seen since the March lows ^[6], and the debate over whether this is a durable cycle or a late-stage bubble is loudest away from institutional desks.

Among retail investors, the skeptics have a standard-bearer: Michael Burry's disclosed short against Micron has become a rallying point on Reddit, where the argument centers on the memory sector's brutal cyclicality rather than its AI tailwind, with commenters noting how often memory names have fallen 30% or more over past cycles. Others draw explicit comparisons to Cisco and the dot-com peak, while a separate thread argues the AI trade is quietly morphing into an energy-demand story as data-center power becomes the binding constraint. The common thread across these community debates is telling: even amid a 300%-plus run in names like Micron, the crowd closest to the trade is arguing about tops, not bottoms.