Plutux
Three joins, one thesis: S&P 500's September rebalance makes AI's power and storage constraints index-eligible insight cover
Markets / EventMU · 000660.KS · 005930.KS11 min read

Three joins, one thesis: S&P 500's September rebalance makes AI's power and storage constraints index-eligible

Bloom Energy, Illumina, and Everpure—the rebranded Pure Storage—join the S&P 500 on Sept. 21, 2026, replacing Molson Coors, Trade Desk, and Builders FirstSource. The mix is a forced-buyer bet on AI's downstream bottlenecks: fuel cells for hyperscaler power, flash storage for AI workloads, plus a return for genomics after a two-year exile. Combined passive demand runs into the tens of billions, with the clearest supply-chain spillovers at NAND suppliers and power utilities.

Published Sep 5, 2026Updated Sep 5, 2026

Bloom Energy market cap

$74.5B

as of Sept. 4, 2026 close; +7.35% on inclusion news

Everpure market cap

$33.1B

NYSE-listed; Q2 FY2027 (Aug 2, 2026)

Illumina market cap

$33.0B

NASDAQ; Q2 FY2026 (Jun 28, 2026)

S&P 500 effective date

Sept. 21, 2026

prior to market open; announced Sept. 4, 2026

On Sept. 4, 2026, after the close, S&P Dow Jones Indices confirmed three additions to the S&P 500 effective prior to the open of trading on Monday, Sept. 21, 2026. The trio's collective market cap of roughly $140 billion forces a measurable reweighting of U.S. large-cap equity flows into AI's infrastructure layer, not its chip layer. Two of the three names—Bloom Energy and Everpure—answer the two constraints that have started binding after compute stopped being the bottleneck. The third—Illumina—is a story about a company earning its way back into the index after a 2024 exit, not about AI at all.

The forced-buyer math: ~$20B of passive demand, weighted toward power

Passively managed assets tracking the S&P 500 sit at roughly $7 trillion, with the Vanguard S&P 500 ETF (VOO) having crossed $1 trillion in assets in June 2026 and SPY and IVV each running above $600 billion. Roughly0.27% of the index's market cap is being reallocated away from Molson Coors, Trade Desk, and Builders FirstSource and toward the three incoming names on Sept. 21. The bulk of that weight lands in Bloom Energy, whose $74.5 billion market cap is more than double the others. A0.85% float-adjusted weight on Bloom translates to well over $40 billion of mechanical buying from index funds and ETFs, with Everpure and Illumina together adding another $25–30 billion.

The mix is unusually concentrated in capex-leverage AI infrastructure. Two of three new names—Bloom Energy and Everpure—sell into the same hyperscaler buying cycle, while Illumina's addition reflects organic earnings recovery, not thematic reweighting. Forced-buyer flows favor the larger name.

Bloom Energy market cap

$74.5B

as of Sept. 4, 2026 close; +7.35% on inclusion news

Everpure market cap

$33.1B

NYSE-listed; Q2 FY2027 (Aug 2, 2026)

Illumina market cap

$33.0B

NASDAQ; Q2 FY2026 (Jun 28, 2026)

S&P 500 effective date

Sept. 21, 2026

prior to market open; announced Sept. 4, 2026

Bloom Energy: power as the new chokepoint

Bloom Energy's case for inclusion is its Q2 2026 results: revenue of $1.065 billion, up 165.5% year over year, with gross margin of 33.4%. The growth has been underwritten by data-center deals, not the legacy utilities and hospitals that anchored the business a year ago. In April 2026, Oracle expanded a master services agreement for up to 2.8 gigawatts of Bloom Energy solid-oxide fuel cells, more than doubling the 1.2 GW already in deployment, and Equinix has continued to extend its 100+ MW relationship. That kind of forward visibility—a multi-gigawatt backlog to a single hyperscaler—is exactly what the S&P 500 committee is screening for.

The fuel-cell pitch is structural: AI workloads need power that can be sited in 90 days, follow load, and bypass the multi-year wait for utility interconnect. Bloom Energy has demonstrated it can deliver an operating system to Oracle in 55 days, ahead of schedule. That operating model aligns directly with hyperscaler build cycles, where the marginal data center is gated by power, not chips.

Bloom Energy quarterly revenue, last six quarters

Inflection in Q4 2025 accelerates through Q2 2026, coincident with Oracle and Equinix data-center deployments.

Unit: $ millions

Q1 2025

$326M revenue

326

Q2 2025

$401M revenue

401

Q3 2025

$519M revenue

519

Q4 2025

$778M revenue

778

Q1 2026

$751M revenue

751

Q2 2026

$1,065M revenue

1,065

Bloom Energy pulls ~$40B of forced buying into a stock with a trailing P/E of 310x, against a forward P/E of 44x—meaning passive flows arrive before any consensus reset on 2027 earnings. The valuation gap is the trade.

Everpure: storage becomes AI's hidden constraint

Everpure—the rebranded Pure Storage, which changed its ticker from PSTG to the single letter P on April 17, 2026—is the data-storage name in the cohort. Q2 FY2027 (quarter ended Aug. 2, 2026) revenue came in at $1.186 billion, up 38% year over year, with gross margin of 68.4%. The growth has been driven by hyperscaler design wins: on Aug. 10, 2026, Everpure announced its second top-five hyperscaler supply agreement, following an earlier win in the year.

The thesis here is not that Everpure is buying more NAND—hyperscalers increasingly source flash directly. Everpure monetizes the storage control plane at 75–85% gross margins: software, replication, and the AI-data-pipeline layer that sits above commodity NAND. That is why Everpure's market cap has expanded to $33.1 billion even as its gross-profit dollars are still well below NetApp's. The market is paying for the AI workload tailwind, not legacy enterprise storage share.

The three additions, side by side
CompanyTickerMkt CapQ2 RevenueQ2 YoY GrowthGross Margin
Bloom EnergyBE (NYSE)$74.5B$1,065M+165.5%33.4%
EverpureP (NYSE)$33.1B$1,186M+37.7%68.4%
IlluminaILMN (NASDAQ)$33.0B$1,159M+9.5%66.5%
The brief framing called this name \"water.\" It is actually data storage—Pure Storage rebranded in 2026. The mix is thereforetwo AI-infrastructure constraints (power, storage) and one unrelated comeback (genomics), not three infrastructure adjacencies.

Illumina: the biology comeback that doesn't fit the AI thesis

Illumina is the analytical outlier. Its addition has nothing to do with AI and everything to do with operating recovery. Q2 FY2026 revenue of $1.16 billion beat the $1.13 billion consensus by ~3%, and non-GAAP EPS of $1.31 was up 10% year over year. Illumina returned to organic growth—6.5% on a constant-currency basis—after a multi-year stretch where the GRAIL overhang and post-pandemic inventory digestion dragged the stock below the S&P 500's market-cap threshold.

The company has used2026 to refinance and simplify the balance sheet. On Aug. 17, 2026, it priced $300 million of 4.95% notes due 2029, using proceeds to retire $500 million of 4.65% notes maturing Sept. 9, 2026. Sequencing consumables revenue, clinical markets, and the early signs of multiomics adoption have all stabilized. The S&P 500 committee is admitting Illumina because its market cap and earnings have finally passed the screen—not because AI needs more sequencers.

Everpure and Illumina will be removed from the S&P MidCap 400, while outgoing S&P 500 members Molson Coors, Trade Desk, and Builders FirstSource will move to the S&P MidCap 400 and S&P SmallCap 600, respectively.

S&P Dow Jones Indices press release, Sept. 4, 2026

The deletions: a thematic rotation away from staples, ads, and housing

The three deletions matter as much as the three additions. Molson Coors is a sub-$15 billion consumer-staples brewer that no longer clears the S&P 500's market-cap threshold; it had been the last beverage name in the index. Trade Desk is an ad-tech company whose share price has lagged as the ad market has re-priced around AI-driven search. Builders FirstSource is a housing supplier caught in the residential construction cycle. The rotation, in other words, is away from slow-growth, rate-sensitive, post-pandemic laggards and toward companies whose top lines are accelerating on AI capex.

On the announcement day, Bloom Energy jumped 7.35% in the regular session and another 6–7% after-hours, while Trade Desk fell 4.37%. The market reads the rotation as the index committee picking winners.

Supply-chain ripple effects: NAND suppliers and power utilities

Two of the three incoming names sell into the same hyperscaler buying cycle. That means their supply chains will see demand pull even before the S&P 500 inclusion date. Everpure buys NAND flash from Samsung Electronics, SK Hynix, Micron, Kioxia, and Western Digital. The top five NAND suppliers posted combined revenue up83.7% quarter-over-quarter in Q1 2026 on supply tightness, and Everpure's second hyperscaler design win locks in multi-quarter volume. The trade is not direct revenue lift—Everpure monetizes software and control, not NAND dollars—but sustained hyperscaler capex signals sustained tightness through2027.

On the power side, Bloom Energy's 2.8-gigawatt Oracle deal validates the entire behind-the-meter generation thesis that has lifted Vistra, Constellation Energy, and NRG Energy. Fuel cells compete with utility-scale gas and nuclear for AI's marginal megawatt, but the implication is the same: every gigawatt of behind-the-meter generation that gets built pulls forward the day hyperscalers need to sign long-dated power purchase agreements with utilities. Eaton sits one tier below in the stack, supplying the power-management gear that every behind-the-meter and grid build-out needs.

  • Micron sells NAND into Everpure's flash arrays; hyperscaler tightness keeps NAND prices elevated into2027.
  • SK Hynix has both NAND and HBM exposure; the Everpure tailwind is a smaller second-derivative than HBM, but is real.
  • Samsung Electronics is the largest NAND supplier; every Everpure design win absorbs incremental Samsung wafer starts.
  • Vistra and Constellation Energy are the grid-scale power counterparties every behind-the-meter fuel cell is partly substituting for.
  • Eaton supplies power-management gear for both data-center builds and behind-the-meter generation.
  • NetApp and PacBio are the most obvious competitive losers: Everpure's hyperscaler wins pressure legacy storage, and Illumina's re-inclusion strengthens its R&D and M&A capacity.

Stocks the September2026 rebalance actually touches

MMicronMU--
--Vol --
-
Bullish
  • Everpure's second hyperscaler design win locks in sustained NAND demand through 2027, supporting already-tight industry pricing.
  • Q1 2026 NAND revenue rose 96.7% QoQ at $5.95 billion; Everpure's volume is incremental to that tightness, not a substitute.
  • Short-term: shares re-rate with continued price discipline; long-term: NAND ASPs remain the swing variable.
0SK Hynix000660.KS--
--Vol --
-
Bullish
  • NAND and HBM exposure means Everpure's tailwind is a smaller second-derivative than HBM, but it compounds with the broader memory upcycle.
  • Korean memory peers will likely track Everpure's hyperscaler commentary for read-through to pricing power.
  • Watch Q3 2026 NAND ASP commentary for confirmation; short-term the AI HBM story still dominates the narrative.
0Samsung Electronics005930.KS--
--Vol --
-
Bullish
  • Largest NAND supplier globally at ~28% market share; Everpure's volume routes through Samsung's fabs at scale.
  • Everpure's 2nd hyperscaler design win locks in incremental wafer-starts for Samsung over the design life of the arrays.
  • Long-term: NAND ASP discipline is the key swing factor; short-term: HBM allocations to NVIDIA still drive the multiple.
VVistraVST--
--Vol --
-
Bullish
  • Grid-scale nuclear and gas counterparties benefit every time a gigawatt of behind-the-meter generation gets announced—even competition is validation for the AI-power thesis.
  • Bloom Energy's 2.8 GW Oracle deal pulls forward the day hyperscalers must sign long-dated PPAs with Vistra.
  • Short-term: tracking AI capex announcements; long-term: data-center load growth is structural regardless of which generation technology wins.
CConstellation EnergyCEG--
--Vol --
-
Bullish
  • Bloom Energy's fuel-cell wins do not displace nuclear PPAs—they complement them by addressing the speed-to-power gap.
  • Oracle's 2.8 GW commitment signals hyperscalers need every available generation technology to hit AI training capacity targets.
  • Watch hyperscaler PPA announcements as the binary short-term catalyst; long-term is locked in by existing reactor fleets and restart timelines.
EEatonETN--
--Vol --
-
Bullish
  • Power-management gear for data-center builds and behind-the-meter generation; benefits whether the megawatt comes from fuel cells or the grid.
  • Bloom Energy's S&P 500 inclusion is a thematic endorsement of the broader electrical-equipment complex that Eaton anchors.
  • Short-term: data-center capex announcements; long-term: electrification capex across utilities, industry, and AI infrastructure.

Plutux is not an investment adviser. Market data and AI-generated analysis are for information and education only, not investment advice. Disclaimer

© Plutux Technology Limited 2026