Intel I226-V Datasheet Specs vs I225-V and I226-LM

The Intel I226-V is a single-port 2.5GbE controller (MAC and PHY in one 7 x 7 mm chip on a PCIe Gen2 x1 lane) that Intel launched in Q2 2022 to replace the I225-V. Compared with the I225-V, it draws less power (1.3 W vs 1.95 W TDP), adds Energy Efficient Ethernet and drops the I225’s inter-packet-gap bug. Throughput is identical.

No spec page mentions the I226’s own PCIe ASPM L1.2 erratum, which keeps link-drop threads coming in 2026. Buy the I226-V for a new appliance; do not rip out a working I225-V to get one.

Spec table: what Intel actually changed

Figures are from Intel ARK for the I226-V and the I225-V, and from the I225/I226 datasheet, revision 2.6. Datasheet power rows are Intel’s reference-card measurements at 3.3 V, 3 m cable, 25 °C. PCI device IDs are from the Linux igc driver header.

The “3.1” is the PCIe spec revision, not the link speed: both parts run one Gen2 lane at 5.0 GT/s, plenty for 2.5 Gbps. TDP is ARK’s headline; the datasheet’s I225 worst case at 100 m and 55 °C is 2,200 mW. Across four ports, ARK’s figures give 7.8 W versus 5.2 W (“1.3W TDP (x4 = ~5W)” in ServeTheHome’s I226 review). Whole-box figures are in the N100 power consumption breakdown.

I226 vs I226-V: which SKU is which

“I226” is the controller family; every chip carries a suffix, and the -V is the client SKU (“PC/Client/Tablet” on ARK). Intel’s I226 series listing has three orderable controllers:

Specs are from ARK for the I226-V, I226-LM and I226-IT; vPro is from Intel’s feature matrix. The driver header also lists an I226-K (3102), an I226-LMvP (5503) and 125F, the ID a blank-NVM chip reports. If pciconf shows re or rge instead of igc, the port is Realtek; see the Realtek 2.5GbE driver explainer.

Where the I226-V wins

Power. The datasheet’s first I226 highlight is “dramatically reduced power consumption in all silicon states (20-50% reduction)”. Cable disconnected is the extreme case, 250 mW down to 17 mW, which matters on a four-port firewall where two ports usually sit empty. The 2.5G idle saving, 885 mW to 762 mW, is what an always-on WAN port sees.

EEE, on the I226 only. The datasheet: “I225 doesn’t support EEE, but I226 does support EEE.” But igc(4) defaults hw.igc.eee_setting to 1, meaning disabled, so an OPNsense or pfSense box never sees the 600 mW idle figure unless you enable EEE, which is the first thing the link-drop checklist tells you to undo. Treat it as a Linux desktop feature, not a firewall feature.

Cable tolerance. Intel lists “improved cable BER performance, especially on longer cables” as an I226 highlight, and the I226-IT holds 2.5G across -40 to 85 °C where the I225-IT stopped at 70 °C. For a -V part the BER point is what counts: marginal Cat 5e is exactly where I225 links renegotiate down to 1 Gbps.

No IPG erratum. Intel’s I225 specification update says first-version I225 silicon needs a minimum receive IPG of 8 bytes, while IEEE 802.3 requires the MAC to tolerate 5 bytes at 2.5 Gbps. Against a link partner that sends tight gaps, throughput “can be as low as 1-10Mbps”. The workarounds were forcing 1 Gbps, or driver 1.0.0.30 with NVM 1.38, which detect the drops and step the link down automatically. Intel fixed it in I225 v2 silicon, which Puget Systems dates to Q3 2020. Every I226 postdates the fix.

Where the I226-V does not win

Throughput. Same MAC, same PCIe lane, same queues: nothing in the spec gives the I226 a speed edge. Nobody should buy an I226 for speed.

ASPM L1.2. This is the I226’s own erratum. The Linux igc commit that works around it says the part advertises PCIe L1.2 support but its exit latency from that state is longer than the packet buffer can absorb under load, so inbound traffic stalls. The FreeBSD port of the fix adds: “Disabling ASPM system-wide (BIOS or OS ASPM policy) does not fix it.” The driver has to clear the L1.2 enable bit in the device’s own PCIe capability, so a driver new enough to carry the workaround is the cure, not a BIOS toggle. That change cites FreeBSD Bugzilla 279245, “igc(4) I226 (and I225) TX hangups”, so the I225 is not immune to the symptom, only to this cause.

Errata shared by both. A July 2024 igc patch cites erratum 9 of the I225/I226 specification update, revision 2.1, “TSN: Packet Transmission Might Cross Qbv Window”, and shrinks the I226 MAC retry buffer from 192 to 88 bytes. A February 2026 patch proposes PCIe link recovery for I225 and I226 after ASPM L0s transitions return 0xFFFFFFFF on register reads; the PCI maintainer replied that it looks like an erratum Intel should document and suggested a quirk disabling L0s instead. Leave PCIe power management off for the NIC.

PXE. The I226 boots EFI only. Irrelevant for OPNsense, occasionally relevant for a lab that netboots old images.

I226-LM vs I226-V: what the suffix changes

The -LM is the commercial SKU with Intel vPro; the -V is the client SKU without it. MAC, PHY, 1.3 W TDP and PCIe lane are the same, and igc(4) claims both device IDs, so for a firewall they are interchangeable. Intel’s feature matrix lists where they differ:

Per the same page, Intel’s teaming and VLAN support needs third-party software anyway. FreeBSD does both in the OS; an OPNsense forum user runs an I226 on FreeBSD 14.3 with LACP and VLANs with “absolutely no problem whatsoever”.

The -LM has one firewall trap: AMT, which Intel’s I226 product page supports “on systems enabled with Intel vPro technology”. An Arch Linux router builder traced DHCP offers failing on the untagged network to “the I226-LM having vPro Essentials enabled on it”. Tagged VLANs worked, and the fix was moving the LAN to the other 2.5GbE port. On an -LM board, keep the -LM port off the LAN that serves DHCP.

Driver support: the version boundaries

Both parts use igc(4), which the FreeBSD manual says “first appeared in FreeBSD 13.1” and which covers every I225 and I226 SKU plus the I220-V and I221-V. It runs 1000 and 2500 Mbps at full duplex only and defaults hw.igc.max_interrupt_rate to 8000.

The boundaries that caught people, from Netgate’s version table: pfSense CE 2.6.0 and Plus 22.01 (February 2022) are FreeBSD 12.3-STABLE builds, and ServeTheHome’s 2022 advice was that until pfSense 2.7 shipped, “the i225-V’s are likely the better option compared to the i226-V”, because 2.6 supported the I225. CE 2.7.0 (June 2023) moved to a FreeBSD 14-CURRENT base and CE 2.8.0 (May 2025) to 15-CURRENT. OPNsense 22.1 (January 2022) moved to FreeBSD 13, and its 22.1.9 changelog already carries an igc fix. Anything current on either platform is fine; the exception is an install image someone burned in 2022.

The I226 datasheet: document 596659

There is no standalone I226 datasheet. Both generations share the Intel Ethernet Controller I225/I226 Datasheet, Intel document 596659, and two revisions circulate as distributor-hosted PDFs: revision 2.6, February 2022, the source of this article’s power figures, and the later revision 2.7.1. Errata live in separate specification updates: the public I225 update, revision 1.2 for the IPG bug, and the combined I225/I226 update (revision 2.1, cited above) for the TSN erratum.

I226-V NVM firmware versions and how to update them

The NVM is the controller’s onboard firmware, separate from the OS driver, and several I226-V link problems were fixed there rather than in igc(4). Protectli says stock firmware on its I226-V units is “typically around NVM 2.17” and that it has validated updates to 2.27 and later. On the OPNsense forum, 2.17 is described as “a problem NVM, dropping connection stuff”.

Intel does not post I226 images publicly. hungvu.tech reports that the update tool goes only to OEMs and system integrators, and Protectli says the .bin files come from Intel via RDC or DevZone. The release notes below are from an unofficial community collection, not checked against Intel’s originals:

Images come in 1 MB and 2 MB variants with different eTrack IDs (80000308 and 80000303 for 2.17), and the image must match the board’s flash chip.

If a box on 2.17 or older still flaps after the link-drop checklist, this is the Linux update path an OpenWrt forum user documented in September 2026:

1. Ask the appliance vendor for an image first; it is the official source.
2. Extract NVMUpdatePackage/I225/I225.tar.gz from Intel’s Ethernet driver pack into a work folder.
3. Run sudo ./nvmupdate64e -i -l inventory.txt and note each port’s ETrackID, which also identifies 1 MB or 2 MB flash.
4. Edit nvmupdate.cfg: NVM IMAGE is the .bin file, EEPID the new eTrack, REPLACES the old one.
5. Run sudo ./nvmupdate64e -u -l -b -c nvmupdate.cfg; -b backs up the current image first.
6. Confirm the new version with ethtool -i.

The forum author’s own warning is “YOU RISK BRICKING YOUR NIC”. Other users in the thread report 2 MB images failing and 1 MB updates stalling at 30%, and hungvu.tech warns that flashing non-V variants “may brick your device”.

What you actually need

• New N100 or N150 appliance: I226-V, for lower draw, longer support and no IPG history. The N100 build guide covers the rest of the spec and firewallcompare’s mini PC list which boxes ship it.
• Existing I225-V box that holds 2.5G: keep it. On v1 silicon with a switch that triggers the IPG bug, force 1 Gbps or get the NVM 1.38 update from the board vendor.
• Flapping I226-V: not a reason to swap to I225. Update the OS so the L1.2 workaround is in the driver, work the link-drop checklist, then check the NVM version.
• Industrial or TSN: I226-IT, the only SKU rated for 2.5G at 85 °C. For a 1 to 4 VLAN firewall the client -V is the right size; pfSense’s hardware requirements cover the CPU and RAM side.

How to check the controller, stepping and NVM on your box

On OPNsense or pfSense:

pciconf -lv | grep -A3 igc
ifconfig igc0 | grep -E 'media|status'
sysctl hw.igc
dmesg | grep -i eeprom

pciconf gives the device ID and revision: 125c is the I226-V (and the ID the FreeBSD L1.2 workaround keys on), 125b the -LM, 15f3 the I225-V. The media line should read 2500Base-T <full-duplex>, sysctl hw.igc should show eee_setting: 1, and dmesg reports the NVM in the form the OPNsense forum quotes, EEPROM V2.25-0 eTrack 0x800003ad.

On Linux:

lspci -nn | grep -i ethernet
sudo lspci -vv -s <bus:dev.fn> | grep -E 'ASPM|L1Sub'
sudo ethtool --show-eee <iface>
ethtool -i <iface>

lspci -vv shows whether L1.2 is enabled on the device itself; a kernel carrying the igc L1.2 commit shows it disabled on an I226 regardless of BIOS. ethtool --show-eee should report no EEE support on an I225 and support on an I226. In ethtool -i, the firmware-version field encodes the NVM: the OpenWrt thread shows 2025:8a6 on NVM 2.25 and 2032:88b6 after flashing 2.32.

Then load it:

iperf3 -c <lan-host> -P 4 -t 60
mtr -n -c 200 <lan-host>

A healthy link holds its rate for 60 s with zero mtr loss. One that drops below 2.5G in ifconfig afterwards points to the IPG bug on I225 v1 silicon, or an L1.2 stall on an I226 whose driver is too old to carry the fix.

The N100 throughput and power calculator shows how NIC choice and enabled services affect the estimated workload envelope.

FAQ

is the i226-v better than the i225-v?

Yes, for any new purchase. The I226-V cuts ARK TDP from 1.95 W to 1.3 W, adds 802.3az EEE, tolerates longer cables and never had the I225’s inter-packet-gap bug, at the same $2.87 ARK price. Throughput is identical, so a working I225-V box gains nothing from a swap.

can i use an i226-lm instead of an i226-v in a firewall?

Yes. OPNsense and pfSense drive both with the same igc(4) driver, and the MAC, PHY and PCIe lane are identical. The -LM adds vPro and AMT manageability, which a firewall never uses and which has interfered with DHCP on at least one router build. Keep the -LM port off your DHCP-serving LAN.

where can i download the intel i226 datasheet?

There is no separate I226 datasheet. Intel covers both chips in the shared I225/I226 datasheet, document 596659. Revision 2.6 (February 2022) and the later revision 2.7.1 circulate as distributor-hosted PDFs, and Intel publishes errata separately in its specification updates. For anything newer, ask the board vendor or Intel.

what is the latest i226-v nvm version?

The newest I226-V image in the community-maintained collection of Intel NVM files is 2.32, which fixes MDI lane-swap polarity handling. Intel does not publish I226 NVM images for public download, so board vendors decide what ships. Protectli reports stock units typically around 2.17 and has validated updates to 2.27 and later.