Polygon's Gigagas Bet Is No Longer Just About Speed

The 100,000-transactions-per-second target dominated Polygon’s June 12, 2025 Gigagas roadmap. But the blueprint was never only a speed project: it also specified roughly five-second finality, smoother fees, fewer chain reorganizations and high uptime—the properties that payment applications and real-world assets need before headline capacity becomes operationally useful.

Fifteen months later, Gigagas is better understood as the implementation of that payments-oriented design. Polygon has widened execution capacity, shortened the time applications wait for irreversible state, rebuilt block production and added recovery mechanisms, then placed a product layer above the chain that handles the parts of money movement a blockchain does not.

The first two milestones shipped on the network then called Polygon PoS, which Polygon now presents as Polygon Chain. Bhilai went live on July 1, 2025, raising the block gas limit from 30 million to 45 million and giving the network stated capacity above 1,000 TPS. The same release dampened block-to-block base-fee changes and modernized the execution stack, so the additional room came with less erratic cost behavior rather than capacity alone.

Heimdall v2 followed on July 10 and reduced deterministic finality from roughly one to two minutes to approximately two to five seconds. A transaction appearing in a block means the producer included it; finality means the network has agreed on a state that applications can treat as irreversible, a distinction reflected in Polygon’s finality model. A wallet can display an incoming transfer before that threshold, but an exchange crediting a balance or a merchant releasing goods may prefer to wait until the transaction is finalized.

That faster chain-level certainty is important, but it is not the same thing as completing a payment. Merchants, exchanges and financial systems need confidence that onchain funds will not disappear in a rollback; identity verification, sanctions screening, foreign-exchange conversion, bank transfers and local payouts operate on separate systems and can continue after blockchain settlement is final. A two-to-five-second protocol guarantee therefore solves one critical segment of the workflow, not the timing of the entire customer journey.

Rio, activated on October 8, addressed the architecture behind both throughput and reliability. It changed how blocks are produced and separated that job more clearly from network-wide validation, while witness-based stateless validation let nodes verify blocks using proofs of the relevant state instead of carrying the same storage and compute burden. That made verification lighter and created an architecture able to support more throughput with less exposure to competing block histories; Polygon attributes stronger claims that Rio effectively eliminated reorganizations to this production model.

Rio enabled the next capacity steps, but it did not independently deliver the ceiling Polygon advertises today. In early 2026, Polygon raised the block gas limit from 60 million to 110 million and claimed capacity above 2,600 TPS. Giugliano, deployed on April 8, then focused on the behavior around that headroom: fee-market parameters became readable onchain, transaction propagation was strengthened for traffic spikes, block information moved through the network sooner to improve confirmations, and execution nodes gained failover paths across Heimdall endpoints.

Those details matter more to a payment operator than another isolated peak-throughput test. Onchain fee parameters make quoting and reconciling transaction costs less dependent on offchain assumptions; stronger propagation reduces the chance that heavy demand leaves nodes working from stale information; endpoint failover removes a single infrastructure dependency during restarts or network disruption. Together, the changes made higher capacity easier to consume as a service with measurable cost and failure behavior.

On June 17, 2026, Polygon configured the chain for a 160-million block gas limit and 1.5-second blocks. Polygon characterizes that configuration as capacity for up to 5,000 payments per second, but the figure is not an independently reproduced benchmark: the announcement does not disclose enough about workload composition, payment gas assumptions, transaction complexity or reproducible methodology to verify it independently. It should also not be treated as interchangeable with every kind of transaction, since a simple stablecoin transfer and a more complex contract interaction consume different amounts of gas.

Capacity is also different from observed use. A PolygonScan snapshot dated September 10, 2026 showed approximately 6.1 million transactions over 24 hours, 73.2 average TPS and 24.3% network utilization. The useful reading is that Polygon has headroom for bursts and future payment flows—not that it continuously processes 5,000 real user payments every second.

Ithaca’s July 30 mainnet upgrade moved the engineering emphasis from adding room to managing failure. It introduced automatic replacement of a stalled block producer, safeguards that reject oversized transactions before they can pull the network off pace, and clearer operational visibility for node operators. Polygon said a producer failure previously could leave the chain without automatic recovery for roughly 15 minutes; for checkout, payroll or treasury software, removing that intervention window is a concrete infrastructure improvement even though it does not raise the throughput headline.

The product logic became explicit on September 9 with Polygon’s Open Money Stack architecture. OMS exposes compliance and identity checks, accounts, custodial or non-custodial wallets, fiat ramps, asset conversion, routing and blockchain settlement through one API, with Polygon Chain as the default settlement rail. Instead of making a payments team integrate separate vendors for customer verification, bank deposits, wallets, liquidity and payouts, the stack is designed to coordinate the handoffs and records across the full flow.

That architecture also puts the chain upgrades in their proper layer. A remittance product might receive fiat, verify the sender, convert funds into a stablecoin, settle value on Polygon, convert again and initiate a bank or cash payout; fast deterministic finality reduces risk in the onchain middle, while fee visibility and failover help the operator price and run that segment predictably. Neither the chain’s two-to-five-second finality nor any broader Polygon settlement estimate guarantees that the receiving bank, compliance provider or payout network finishes at the same speed.

OMS should still be treated as an early product as of September 10, 2026. It entered technical preview in June, and Polygon’s current site continues to ask users to request early access; feature coverage also varies by network, corridor and region. The September architecture shows a more complete infrastructure proposition, but it does not establish general availability or broad production performance across the jurisdictions and payment methods OMS intends to cover.

PYUSD offers a useful external check on the institutional framing. Polygon announced native support in July 2026, and Paxos documentation lists a PYUSD contract on Polygon PoS mainnet. The relevant signal is not a longer partner roll call, but that a regulated payment asset has reached the settlement rail Polygon has spent the past year hardening and is intended to be usable inside the wider OMS workflow.

Gigagas, then, has not pivoted from speed toward reliability. Polygon is implementing the payments design already embedded in the original roadmap, where throughput is one requirement alongside irreversible settlement, predictable costs, lighter verification and graceful recovery—and where the chain becomes more useful when those properties are packaged with compliance, fiat access and payouts. Polygon states that 160M-gas blocks provide capacity for up to 5,000 payments per second; the original roadmap places 100,000 TPS in “2026 and beyond,” but Polygon has not provided a confirmed delivery date.